Abstract
Laser powder bed fusion (LPBF) technic is suitable for manufacturing metal parts with complex shapes and internal channels. The energy input and post heat treatment have great influence on the formability, microstructure and properties of the metallic materials. In this work, Hastelloy X superalloy is fabricated by LPBF with a variation of laser power, and then post-treated by hot isostatic pressing and solution treatment. Influences of laser power and heat treatment on microstructure features of as-built sample are evaluated. The correlation among laser power, microstructure features and tensile properties of heat-treated samples are revealed. The results reveal that average grain size, aspect ratio, texture intensity increases while the proportion of high-angle grain boundaries decreases with the increase of laser power. Precipitates of Cr-rich carbides along grain boundary are observed after post heat treatment. Post-treatment promotes the twinned recrystallization, grain coarsening and texture degradation. Yield strength increases while tensile strength decreases with the increase in laser power. The higher vertical yield strength of sample with laser power of 320 W could be attributed to its larger aspect ratio and more twin boundaries. 280 W is the preferred laser power to obtain the best overall tensile properties. This study is helpful for the exploration of optimal LPBF process and post heat treatment, by which Ni-based superalloys with a good balance between strength and ductility can be produced.
Export citation and abstract BibTeX RIS
Original content from this work may be used under the terms of the Creative Commons Attribution 4.0 licence. Any further distribution of this work must maintain attribution to the author(s) and the title of the work, journal citation and DOI.
1. Introduction
As one of the most popular additive manufacturing (AM) technics for metals, laser powder bed fusion (LPBF) is used to fabricate metallic components with complex shapes and internal channels. The principle of LPBF is that the powder bed is selectively melted by concentrated laser along the certain route, and quickly bonded, so as to form parts layer-by-layer [1, 2]. Compared to other metallic AM techniques, LPBF has the advantage of high forming precision and good surface finishing [3]. Ni-based superalloys such as Inconel 625, Inconel 718 and Hastelloy X, are the mature alloys processed by LPBF in terms of research, production and adoption. Among them, Hastelloy X (HX) is a solid solution strengthened alloy which exhibits high oxidation resistance, mechanical properties and formability at elevated temperatures up to 1000 °C. It has been widely applied in gas turbine engine components and in the chemical processing industry [4, 5].
However, as a rapid prototyping method, the small molten pools and the extremely fast cooling rate contribute to the nonuniform temperature field and steep temperature gradient during LPBF process, and the ensuing large thermal stress leads to the tendency of thermal cracking and deformation in the formed parts [6, 7]. The problems such as microstructure inhomogeneity, solidification defects and anisotropy in mechanical properties hinder the practical applications [8–10]. It is known that the energy density of heat input has a crucial effect on the quality of the formed parts since they determine the heating/cooling rate, the thermal history and the shape of melt pool. For a particular material, the energy density involves the specific process parameters of laser power, scan speed, layer thickness and hatch spacing or beam diameter [11]. These parameters need to be coordinated to control the microstructure, texture, porosity, surface quality, defects and mechanical properties of the formed parts [12].
In recent years, the influence of process parameters on microstructure and property of Ni-based superalloy parts formed by LPBF has attracted extensive attention [13–15]. Some studies have shown that the increased energy input via increasing laser power, reducing scan speed, layer thickness and hatch spacing, would result in stronger crystallographic textures and inhomogeneous microstructure [16, 17]. Huang et al [13] reported the orientation dependent microstructure and mechanical behavior of HX and revealed the governing mechanism for the strength anisotropy was the columnar grain structure. The work of Liu et al [15] demonstrated the difference in Taylor factor distribution caused by texture anisotropy is the main reason for the difference in tensile properties of samples built at different orientations and energy densities. And it revealed that texture enhances the anisotropy of yield strength, but has little effect on ultimate strength and plasticity. It was proposed by Pilgar et al [18] that the differences in the mechanical response observed for different printing direction and thickness can be fully attributed to the changes in the texture and grain aspect ratios induced by the LPBF process. It was reported by Keller et al [19] that samples elaborated with the rescanning strategy exhibit weaker crystallographic texture and larger ductility compared to conventional LPBF. Compared with the solution treated sample, the second scan (with a 60% laser power of the first scan) increases the yield stress by 2.5 times without ductility reduction.
Besides the optimization of process parameters, proper post-treatments such as solution treatment (ST) and hot isostatic pressing (HIP) are necessary for the important metallic parts [20, 21]. It was reported that the HX samples heated at 1177 °C showed less carbide formation due to their more homogeneous starting microstructure, which gave rise to a tensile ductility increase as compared to the as-built and 800 °C heat-treated samples [22]. It was found that solution treatment at 1177 °C for different time would lead to different degrees of static recrystallization in the LPBF fabricated HX [23]. By controlling cellular subgrain microstructure through subsequent homogenization and double-aging heat treatment, Zhao et al achieved a simultaneous improvements of strength and ductility of Inconel 718 [24]. The high temperature and pressure during HIP process promote the closure of internal microcracks and gas-free pores, recrystallization and precipitation, which reduce stress concentration and release residual stress, leading to the significant improvements in elongation and fatigue life [25, 26]. However, the influence of HIP on the ultimate tensile stress of LPBF fabricated HX is inconsistent. Although there have been many researches, the understanding of the process-microstructure-property relationship of LPBF fabricated Ni-based superalloys is still limited.
In this paper, HX superalloy is fabricated by LPBF with a variation of laser power, then post-treated by hot isostatic pressing and solution treatment (HIP+ST). The differences in solidification microstructure, texture intensity, grain size and grain boundary feature of as-built samples with different laser powers are evaluated. Effects of heat treatment on cracks and microstructure are investigated. In addition, influences of laser power on microstructure and tensile properties of heat-treated samples are evaluated. The correlation among laser power, microstructure features and tensile properties is deeply discussed. This study is helpful for the exploration of optimal LPBF process and post heat treatment, by which Ni-based superalloys with a good balance between strength and ductility can be produced.
2. Experimental procedure
A BLT-S210 LPBF machine made by Bright Laser Technologies company was used to conduct the fabrication. The particle sizes of HX powder ranges from 15 to 53 μm, and the chemical compositions is shown in table 1. Cubes (10 × 10 × 10 mm3) and vertical barbell-like samples (70 × Φ15 mm3) were built for microstructure observation and tensile experiment with the process parameters listed in table 2. The energy input can be quantified by volumetric energy density (VED), which is defined as equation (1),
where P, v, H and t refers to laser power, scan speed, hatch spacing and layer thickness, respectively. Therefore, the VED is 51.9, 55.6 and 59.3 J mm−3 corresponding to the laser power of 280, 300 and 320 W, respectively. As-built samples were subjected to hot isostatic pressing (1180 °C/180 MPa/3 h) followed by the solution treatment 1180 °C/1 h/forced argon cooling).
Table 1. Chemical compositions of Hastelloy X powder (wt%).
| Ni | Cr | Fe | Mo | Co | W | Al | C | Mn | O | N |
|---|---|---|---|---|---|---|---|---|---|---|
| Bal. | 21.60 | 18.55 | 9.44 | 1.93 | 0.30 | 0.09 | 0.07 | 0.01 | 0.01 | 0.01 |
Table 2. LPBF experimental process parameters.
| Preheating | Beam diameter | Layer thickness | Laser power | Hatch spacing | Scan speed | Rotation angle |
|---|---|---|---|---|---|---|
| 100 °C | 100 μm | 60 μm | 280/300/320 W | 90 μm | 1000 mm s−1 | 67° |
A Zeiss Sigma 500 scanning electron microscope (SEM) equipped with an Oxford NordlysMax3 electron backscatter diffractometer (EBSD) was used to observe the microstructure, grain orientation distribution and grain size. The vertical section of cube sample (parallel to the building direction) was ground, and then mechanically polished with oxide polishing suspension (OPS) solution for the EBSD observation. The EBSD field of view was 1500 μm × 1120 μm, with the step size of 2.5 μm. The post-data processing was performed by the Oxford Instrument HKL-Channel 5 software. Misorientation angles of adjacent grains between 2° and 15° were classified as low-angle grain boundaries (LAGBs) and those above 15° were classified as high-angle grain boundaries (HAGBs). Before microstructure characterization by the secondary electron (SE) detector in SEM, samples need to be corroded with the etching solution (HCl: HNO3 = 3:1). The micro-area chemical composition was determined by the Oxford X-MaxN x-ray energy dispersive spectrometer (EDS).
The barbell-like tensile specimens were machined according to ASTM-E8M standard, with the geometry and dimension shown in figure 1. The tensile properties were tested at a loading speed of 1 mm min−1 on a SENS CMT-5105 electronic universal testing machine. Three measurements were taken to calculate the average value of tensile properties.
Figure 1. Schematic diagram for tensile specimen dimension.
Download figure:
Standard image High-resolution image3. Results and discussions
3.1. Effect of laser power on solidification microstructures
The inverse pole figure (IPF) coloring maps of as-built HX with different laser power are shown in figure 2. The as-built microstructure is characterized by the elongated grains and fiber texture along the building direction. It reveals the preferred grain orientation that the crystallographic direction [001] of most grains is parallel to building direction. The large number of black fine lines in the elongated grains reveal the high density of LAGBs inside. Based on grain detection, the average aspect ratio of at-built grains is calculated to be 4.5, 5.0 and 5.2 for figures 2(a)–(c), respectively, which shows an increasing trend with the increase of laser power. The increased laser power (energy density) contributes to the larger melt pools and the lower cooling rate, promoting the epitaxial growth of columnar crystals along the long axis.
Figure 2. IPF coloring maps of as-built HX with laser power of 280 W (a), 300 W (b) and 320 W (c).
Download figure:
Standard image High-resolution imageFigure 3 depicts pole figures of the crystallographic plane (001) for as-built HX. A strong crystallographic texture with maximal density poles larger than 5.7 multiple random distribution (MRD) is shown in these samples. The pole distribution feature is close to the Goss texture of {110}〈001〉, which is usually formed after the recrystallization of deformed metals with cubic lattice [27]. This Goss texture component is related to the specific heat flux direction and associated solidification, and has been already characterized in Ni-based superalloys fabricated by LPBF and laser directed energy deposition [19, 28]. Moreover, it is found that the texture intensity increases with the increase of laser power.
Figure 3. Pole figures of the crystallographic plane (001) for as-built HX with laser power of 280 W (a), 300 W (b) and 320 W (c). ⨂ denotes the {110}〈001〉 Goss texture component.
Download figure:
Standard image High-resolution imageFigure 4 depicts the grain size distribution and grain misorientation distribution of as-built HX. The grain size is expressed by equivalent circular grain diameter. As shown in figures 4(a)–(c), the grains with diameter below 20 μm occupy the largest proportion. The very fast solidification rate of metals liquid in molten pool accounts for the formation of fine grains. The number of grains shows a downward trend with the increase of grain size, and very few grains have a diameter larger than 200 μm. As shown in figures 4(d)–(f), the proportions of misorientation angle below 3° are more than 25% and LAGBs are dominant in as-built samples.
Figure 4. Grain size distribution and grain misorientation distribution of as-built HX with laser power of 280 W (a), (d), 300 W (b), (e) and 320 W (c), (f).
Download figure:
Standard image High-resolution imageThe average grain diameter (AGD) is calculated to be 34.8, 37.8 and 41.4 μm, respectively. Correspondingly, the percentage of HAGBs is 41.3%, 39.0% and 36.2%. A significant relevancy was found that the AGD increases while the proportion of HAGBs decreases with the increase of laser power. The increased laser power (energy density) contributes to the larger melt pools and the lower cooling rate, leading to a coarser grain structure. The increased frequency of LAGBs with the increasing laser power indicates the higher density of dislocations, which is consistent with the enhanced texture intensity discussed above [29]. It is found that the higher energy input promotes the growth of grains and the proliferation of dislocations.
3.2. Effects of heat treatment on cracks and microstructure
Figure 5 gives the optical microscopy images of as-built and HIP+ST samples. A small number of cracks and pores are present in as-built samples. The significant temperature gradient, rapid cooling rate of ∼105 K s−1, intergranular carbides and high thermal residual stress accumulation promote the formation of hot cracks along the grain boundaries and across several pre-solidified layers in LPBF fabricated HX [2, 30]. Pores includes gas pores and gas-free pores which is related to the gas in the powder and the process parameters, respectively. The existence of cracks and pores will pose a serious threat to the fatigue performance of the product.
Figure 5. Optical microscopy images of the as-built HX with laser power of 280 W (a), 300 W (b), 320 W (c) and HIP+ST sample with laser power of 300 W (d). Cracks and minute pores are highlighted with white and black arrows, respectively.
Download figure:
Standard image High-resolution imageIn the field of view in figure 5(a), there are five microcracks between 15 and 50 μm in length and more than a dozen holes of varying sizes. The number of pores increases significantly in as-built HX with the highest laser power of 320 W (figure 5(c)). However, the number of cracks and pores in the present work are quite small as compared to the LPBF HX in refs. [26, 30]. It has been reported that the optimal energy density range yielding the highest relative density for the HX is between 40 and 80 J mm−3 [12]. The result indicates that process parameters in the present work, with a VED range of 50 to 60 J mm−3, are preferable for the LPBF of HX. It can be observed in figure 5(d) that cracks no longer exist in the HIP+ST sample while a few pores still present. HIP is useful in closing internal microcracks, but has limited effect on the elimination of gas pores. Moreover, the gas pores closed by HIP can reopen and grow during subsequent heat treatment [31].
Figure 6(a) gives the SEM image of as-built HX with laser power of 300 W. The microstructure is characterized by the very fine cellular and columnar dendrites. The morphology of grain and internal substructure is determined by the temperature gradient and growth rate at the front edge of the solid/liquid interface. The ultrahigh solidification rate and large thermal gradient in the LPBF molten pool limit the formation and growth of secondary dendrites, most of metals solidify in the cellular mode with complex structural morphology. Temperature gradients and growth rates are different throughout the molten pool. In general, the temperature gradient at the bottom of molten pool is large, which is conducive to the formation of columnar crystals. The intrinsic cyclic thermal treatment during the layer-by-layer fabrication promotes the epitaxial growth of columnar grains, and results in the formation of strong texture along building direction [17].
Figure 6. SEM images of the as-built (a) and HIP+ST (b), (c) HX with laser power of 300 W. The element distribution maps of (c) are shown in (d).
Download figure:
Standard image High-resolution imageAs shown is figure 6(b), after HIP+ST treatment, the microstructure becomes smooth with clear boundaries and these is almost no precipitate remaining in the grains. A continuous thin film composed of precipitates along grain boundary can be observed in the high magnification image (figure 6(c)). The element distribution maps in figure 6(d) and table 3 evidence that the elongated submicron precipitates at grain boundary are Cr-rich carbides, which has been identified to be M23C6 precipitates in the similar research [32]. It is inferred that a complete dissolution of the carbides has occurred during solution treatment, whereas the subsequent argon cooling is insufficient to inhibit the re-precipitation of intergranular carbides during cooling.
Table 3. Chemical compositions (wt.%) of each location in figure 6(c).
| Location | Ni | Cr | Fe | Mo | Co | W | C |
|---|---|---|---|---|---|---|---|
| 1 | 47.9 | 21.8 | 18.5 | 9.0 | 2.1 | 0.5 | 0.2 |
| 2 | 33.9 | 40.4 | 15.6 | 6.9 | 1.5 | 0.8 | 0.9 |
| 3 | 31.0 | 36.9 | 14.6 | 13.4 | 1.8 | 1.0 | 1.3 |
| 4 | 29.3 | 41.4 | 13.9 | 11.6 | 1.4 | 0.8 | 1.6 |
3.3. Influences of laser power on as-heat-treated microstructure and tensile properties
Figure 7 gives the IPF coloring maps of HIP+ST samples with different laser power. The microstructure becomes complex with diverse grain morphology including twins, deformed grains and a little equiaxed grains. The changed grain morphology indicates the occurrence of static recrystallization during heat treatment, despite the microstructure is quite different to the deformed-recrystallized microstructure composed of equiaxed grains. The high distortion energy in the as-built HX and the high heat treatment temperature promote the occurrence of static recrystallization. However, a few elongated grains remain in HIP+ST samples. Based on grain detection, the average aspect ratio of grains in figures 7(a)–(c) are calculated to be 2.0, 2.1 and 2.2, respectively, showing a positive correlation with laser power. Figure 8 depicts pole figures of the crystallographic plane (001) for HIP+ST samples, which evidences that the orientation intensity of grains decreases obviously. The weakening of microstructural anisotropy by heat treating will greatly reduce the anisotropy on mechanical behaviors [33, 34].
Figure 7. IPF coloring maps of HIP+ST samples with laser power of 280 W (a), 300 W (b) and 320 W (c).
Download figure:
Standard image High-resolution image
Figure 8. Pole figure of the crystallographic plane (001) for HIP+ST samples with laser power of 280 W (a), 300 W (b) and 320 W (c).
Download figure:
Standard image High-resolution imageFigure 9 depicts the grain size distribution and grain misorientation distribution of HIP+ST samples. As shown in figures 9(a)–(c), the proportion of fine grains is the highest, and the maximum grain size is further increased relative to the as-built state. The AGD of is calculated to be 43.1, 48.6 and 50.4 μm, increasing by 24%, 29% and 22% respectively, as compared to the corresponding as-built sample. The grain size of HIP+ST sample also increases with the increasing laser power, maintaining the same changing regular as the as-built state. As shown in figures 9(d)–(f), heat treatment decreases the LAGBs dramatically. The dominant misorientation angle is in the vicinity of 60° which can be identified as Σ3 twin boundaries, indicating the formation of numerous annealing twins [35]. The proportion of Σ3 boundaries is 38.3%, 39.8% and 42.0%, respectively. It is deduced that the distortion energy and residual stress of as-built samples increase with the increased laser power, facilitating the formation of more annealing twins. The percentage of HAGBs increases to 86.4%, 84.4% and 85.0%, respectively. The elimination of LAGBs by heat treating indicates the decrease of dislocation density and removal of internal stress in as-built samples. It is noticed that the difference in grain misorientation distribution induced by varying energy inputs (figure 4) has been wiped out by heat treating.
Figure 9. Grain size distribution and grain misorientation distribution of HIP+ST samples with laser power of 280 W (a), (d), 300 W (b), (e) and 320 W (c), (f).
Download figure:
Standard image High-resolution imageThe recrystallization map base on the grain orientation spread of HIP+ST samples with different laser powers are shown in figure 10, where grain boundaries are displayed in black lines and the recrystallized grains, subgrains and deformed grains are distinguished by blue, yellow and red, respectively. It is observed that the deformed grains are generally elongated with a high density of LAGB inside, same as the those in as-built state. Considering the fact that Σ3 twin boundaries are dominant in recrystallized region, it is realized that recrystallization during HIP+ST is occurred mainly by twinning. As listed in table 4, the proportion of recrystallized region are more than 63% while the proportion of deformed grains are less than 10%. The remaining portion represents subgrains, which retain their original shapes during recovery process, but with reduced dislocation density, lattice distortion and internal stress. Reasons for the incomplete static recrystallization could involve the insufficient stored energy in some grains and the inadequate annealing time. In addition, according to researches [23, 36], some pre-existing fine particles such as Al-Ti-O act as a barrier for grain boundary migration and slow down the progress of recrystallization. It is noticed that after HIP+ST, the recrystallization degrees for samples with different laser powers are similar.
Figure 10. Recrystallization map of HIP+ST samples with laser power of 280 W (a), 300 W (b) and 320 W (c). Bule: recrystallized grains, yellow: subgrains, red: deformed grains. The minimum misorientation angle to separate subgrains and deformed grains is set to be 1.5° and 7.5°, respectively.
Download figure:
Standard image High-resolution imageTable 4. Recrystallized percentage in figure10.
| 280 W | 300 W | 320 W | |
|---|---|---|---|
| Recrystallized grains | 66.8 | 63.9 | 69.5 |
| Subgrains | 26.6 | 26.2 | 23.7 |
| Deformed grains | 6.6 | 9.9 | 6.8 |
Tensile properties of HIP+ST samples are given in figure 11. The yield strength is 309, 318.3 and 325 MPa for sample with laser power of 280, 300 and 320 W, respectively, which is positively correlated with laser power. However, the contrary regular was observed for tensile strength which decreases from 702 to 677 and 666.7 MPa. Considering the positive correlation between AGD and yield strength, it is found that Hall-Petch effect that yield strength decreases with coarsening of grains is not suitable to explain the increase of yield strength in this work. This could be understood that grains in HIP+ST samples are not equiaxed but with diverse shapes, the equivalent AGD can not reflect the behaviors of actual grains and grain boundaries against plastic deformation. It is speculated that factors such as grain aspect ratios, annealing twins, remaining texture effect and recrystallized fraction play important roles in determining the yield strength of HIP+ST samples. Specifically, the larger aspect ratio and proportion of twin boundaries in samples with higher laser power contribute to improved vertical yield strength. It can be observed from figure 10 (especially figure 10(c), 320 W) that after HIP+ST treatment, deformed grains and subgrains still keep the growth along vertical direction. This part of grains has a large aspect ratio, and still maintains a relatively high density of dislocations along major axis. It provides more resistance against deformation, so as to improve the vertical tensile property. Twin boundaries can serve as barriers to dislocation movement, similar to grain boundaries. When you have a higher proportion of twin boundaries, it can further impede dislocation motion, thus contributing to improved yield strength.
Figure 11. Tensile properties of vertical samples with different laser power after HIP+ST.
Download figure:
Standard image High-resolution image
Based on the difference between tensile strength and yield strength, the strain-hardening effect is found gradually decreases as the laser power increases. Strain hardening is affected by grain size and distribution. In general, fine, evenly distributed grains usually have a higher strain hardening effect and ductility. In this experiment, the grain size and aspect ratio increase as laser power increases. Sample with the minimum laser power exhibits the higher strain hardening effect and elongation, which is consistent with the above views. Therefore, it can be understood that the increase in laser power leads to a larger grain size, which reduces the strain hardening effect. The elongation reaches around 45% thanks to the elimination of texture and closure of micro-cracks via HIP. In this work, 280 W is the preferred laser power to obtain the best overall tensile properties. This is of guiding significance for LPBF parameter optimization and performance prediction of Hastelloy X in the future.
4. Conclusions
In this work, effects of laser power and heat treatment on microstructure features of as-built sample are evaluated. The correlation among laser power, microstructure features and tensile properties of heat-treated samples are revealed. The main conclusions are listed as follows.
- 1.The as-built microstructure is characterized by the elongated grains and Goss texture of {110}〈001〉 along the building direction. The average grain size, aspect ratio and texture intensity increase as the laser power increases, while the proportion of HAGBs decreases. The increased laser power (energy density) promotes the epitaxial growth of columnar crystals.
- 2.HIP allows to close all cracks in as-built parts but with gas pores remaining. Precipitates of Cr-rich carbides along grain boundary are observed after ST due to the insufficient cooling rate.
- 3.HIP+ST contributes to the twinned recrystallization, grain coarsening and texture degradation. Most of the as-built grains are recrystallized via twinning, so as the heat-treated microstructures are composed of twinned recrystallization grains, deformed grains without and with high density dislocation, and a little equiaxed grains. The initial grain size of the LPBF microstructure plays an important role in recrystallization. After HIP+ST, a more than 22% increase in grain diameter was observed along with a significant reduction in LAGBs.
- 4.Yield strength increases while tensile strength decreases with the increase in laser power, from which the strain-hardening effect is found gradually decreases. It is considered that the higher aspect ratio and proportion of twin boundaries contribute to the improved vertical yield strength. The increase in laser power leads to a larger grain size, which could be responsible for the reduced strain hardening effect. 280 W is the preferred laser power to obtain the best overall tensile properties.
Data availability statement
All data that support the findings of this study are included within the article (and any supplementary files).
Funding
This research was funded by the National Key R&D Program of China (No. 2022YFB4600800).